Real-time people and vehicle detection from UAV imagery

Gaszczak, Anna; Breckon, Toby P.; Han, Jungong

doi:10.1117/12.876663

Cited by 159 publications

(135 citation statements)

References 18 publications

Supporting

Mentioning

128

Contrasting

Unclassified

Order By: Relevance

“…Additionally, reasoning between different, correctly classified objects becomes more difficult since texture information is rare. In [7], thermal as well as visible-light images are combined to detect cars and humans from a UAV viewpoint. In [8], thermal cameras have been applied to track pedestrians from an upfront viewpoint for night driving using a fusion of the hyper permutation network, a hierarchical contour matching algorithm and a cascaded classifier, respectively.…”

Section: B From Visible-light To Thermal Image Sequencesmentioning

confidence: 99%

People detection and tracking from aerial thermal views

Portmann

Lynen

Chli

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

135

View full text Add to dashboard Cite

Abstract-Detection and tracking of people in visible-light images has been subject to extensive research in the past decades with applications ranging from surveillance to searchand-rescue. Following the growing availability of thermal cameras and the distinctive thermal signature of humans, research effort has been focusing on developing people detection and tracking methodologies applicable to this sensing modality. However, a plethora of challenges arise on the transition from visible-light to thermal images, especially with the recent trend of employing thermal cameras onboard aerial platforms (e.g. in search-and-rescue research) capturing oblique views of the scenery. This paper presents a new, publicly available dataset of annotated thermal image sequences, posing a multitude of challenges for people detection and tracking. Moreover, we propose a new particle filter based framework for tracking people in aerial thermal images. Finally, we evaluate the performance of this pipeline on our dataset, incorporating a selection of relevant, state-of-the-art methods and present a comprehensive discussion of the merits spawning from our study.

show abstract

Section: B From Visible-light To Thermal Image Sequencesmentioning

confidence: 99%

People detection and tracking from aerial thermal views

Portmann

Lynen

Chli

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

135

View full text Add to dashboard Cite

show abstract

“…The thermal hot spot points were analysed using watershed algorithm to generate the seed point. The seed point is the reference for the detection of the vehicles [2].…”

Section: Related Workmentioning

confidence: 99%

Real time Traffic Jam Detection from UAV Using Cascade Haar

S¹,

Saad²

2016

Fourth International Conference on Advances in Computing, Communication and Information Technology CCIT- 2016

View full text Add to dashboard Cite

Abstract-This paper contains an in depth work carried out in the area of real-time vehicle detection from an Unmanned Ariel Vehicle (UAV) to detect the traffic jam by counts the number of cars in the roadway. The optical camera located inside the UAV takes images from a particular angle and analysis is performed on the image to detect the vehicle (static or moving). To obtain the accurate and quick result, the analysis should be in ground station. This paper presents a highly reliable technique to detect many different types of vehicles from images having variant backgrounds. This technique used Haar classifiers to perform this real-time application. The Haar classifier was trained with multiple features to accomplish the task. The task was successfully achieved with minimum false positive rate.

show abstract

“…Nowadays little work has been done on detecting human from a UAV. Gszczak et al (2011) proposed to use both thermal and visible imagery to better detect people and cars, features extracted on thermal and visible imagery are fused together boosting the confidence level of detection. Indeed, the thermal camera is used for extracting Haar-like features while the optical camera is used for a contour shape analysis as a secondary confirmation to better confirm the detection.…”

Section: Existing Work For Detecting Human From the Airmentioning

confidence: 99%

How to improve the HOG detector in the UAV context

Blondel

Potelle

Pégard

et al. 2013

IFAC Proceedings Volumes

View full text Add to dashboard Cite

The well known HOG (Histogram of Oriented Gradients) of Dalal and Triggs is commonly used for pedestrian detection from 2d moving embedded cameras (driving assistance) or static cameras (video surveillance). In this paper we show how to use and improve the HOG detector in the UAV context. In order to increase the elevation angular robustness we propose to use a more appropriate training dataset and sliding windows. We show results on synthetic images.

show abstract

Real-time people and vehicle detection from UAV imagery

Cited by 159 publications

References 18 publications

People detection and tracking from aerial thermal views

People detection and tracking from aerial thermal views

Real time Traffic Jam Detection from UAV Using Cascade Haar

How to improve the HOG detector in the UAV context

Contact Info

Product

Resources

About